Forums

compression speed

Started by hyperbob April 25, 2004
Hello,

I am wondering how much computation it takes to compress audio. This
information is a little hard to come by.

The general situation is this: I am working on a program that will
record either from the sound cards of several computers or from some
card that provides several raw audio channels to be recorded from. I
don't know how to estimate if I can compress all the audio in real
time.

How long does it usually take to compress one channel of audio
(particularly, is it usually real-time)? Will real-time compression
take all of the computer's resources, or could other computations be
done at the same time? When the same computer needs to encode several
channels of audio, could it be reasonably done in real-time?

I am thinking particularly about free or open-source codecs. Do they
tend to be distinctly different in performance than the patented ones?

Thanks a lot,

Bob.
A GSM cellphone Enhanced Full-Rate (EFR) vocoder uses around 20 MIPS

--Anthony


hyperbob wrote:

> Hello, > > I am wondering how much computation it takes to compress audio. This > information is a little hard to come by.
What do you mean by compress? It has at least 2 very distinct meanings for audio. Regards, Dan. -- And on the evening of the first day, the lord said.... LX1, Go! And there was light. The email address *IS* valid, do not remove the spamblock.
Audio or Speech?? You mentioned audio, but just wanted to make sure.
The codecs are very different for the two. You are getting people
confused right now :-)

Regards
Piyush


hyperbob@walla.com (hyperbob) wrote in message news:<c0c6c733.0404250314.f23b262@posting.google.com>...
> Hello, > > I am wondering how much computation it takes to compress audio. This > information is a little hard to come by. > > The general situation is this: I am working on a program that will > record either from the sound cards of several computers or from some > card that provides several raw audio channels to be recorded from. I > don't know how to estimate if I can compress all the audio in real > time. > > How long does it usually take to compress one channel of audio > (particularly, is it usually real-time)? Will real-time compression > take all of the computer's resources, or could other computations be > done at the same time? When the same computer needs to encode several > channels of audio, could it be reasonably done in real-time? > > I am thinking particularly about free or open-source codecs. Do they > tend to be distinctly different in performance than the patented ones? > > Thanks a lot, > > Bob.
I was refering to speech compression (of the space-saving sort). Sorry
about the confusion :-).

Thanks,

Bob.

piyushkaul@yahoo.com (Piyush Kaul) wrote in message news:<2e20232a.0404252239.64d78ecb@posting.google.com>...
> Audio or Speech?? You mentioned audio, but just wanted to make sure. > The codecs are very different for the two. You are getting people > confused right now :-) > > Regards > Piyush > > > hyperbob@walla.com (hyperbob) wrote in message news:<c0c6c733.0404250314.f23b262@posting.google.com>... > > Hello, > > > > I am wondering how much computation it takes to compress audio. This > > information is a little hard to come by. > > > > The general situation is this: I am working on a program that will > > record either from the sound cards of several computers or from some > > card that provides several raw audio channels to be recorded from. I > > don't know how to estimate if I can compress all the audio in real > > time. > > > > How long does it usually take to compress one channel of audio > > (particularly, is it usually real-time)? Will real-time compression > > take all of the computer's resources, or could other computations be > > done at the same time? When the same computer needs to encode several > > channels of audio, could it be reasonably done in real-time? > > > > I am thinking particularly about free or open-source codecs. Do they > > tend to be distinctly different in performance than the patented ones? > > > > Thanks a lot, > > > > Bob.
I am assuming you are going to run this on a Pentium PC. I remember
using the (encoder/decoder) of  floating-point reference
implementation implimentation of G.729 on PC. It would take a less
than 10% cpu on 400 MHz Pentium III for a single channel. Use the
floating point implementations if you can. It is faster.  Other CELP
based codecs like GSM-AMR/G.723.1 etc would also take processing in
the same range. Hope that is the idea you were looking for. As for
"how long", most such speech codecs have a frame lengths of ~10-20 ms,
and the algorithmic delay may span ~10-50 ms.

Regards
Piyush


hyperbob@walla.com (hyperbob) wrote in message news:<c0c6c733.0405020232.6083763b@posting.google.com>...
> I was refering to speech compression (of the space-saving sort). Sorry > about the confusion :-). > > Thanks, > > Bob. > > piyushkaul@yahoo.com (Piyush Kaul) wrote in message news:<2e20232a.0404252239.64d78ecb@posting.google.com>... > > Audio or Speech?? You mentioned audio, but just wanted to make sure. > > The codecs are very different for the two. You are getting people > > confused right now :-) > > > > Regards > > Piyush > > > > > > hyperbob@walla.com (hyperbob) wrote in message news:<c0c6c733.0404250314.f23b262@posting.google.com>... > > > Hello, > > > > > > I am wondering how much computation it takes to compress audio. This > > > information is a little hard to come by. > > > > > > The general situation is this: I am working on a program that will > > > record either from the sound cards of several computers or from some > > > card that provides several raw audio channels to be recorded from. I > > > don't know how to estimate if I can compress all the audio in real > > > time. > > > > > > How long does it usually take to compress one channel of audio > > > (particularly, is it usually real-time)? Will real-time compression > > > take all of the computer's resources, or could other computations be > > > done at the same time? When the same computer needs to encode several > > > channels of audio, could it be reasonably done in real-time? > > > > > > I am thinking particularly about free or open-source codecs. Do they > > > tend to be distinctly different in performance than the patented ones? > > > > > > Thanks a lot, > > > > > > Bob.